So, it turns out what we consider human intelligence is just a shitty alogrythm implemented in meat.
I wasn't expecting basic math with the right kinds of parameters and right kinds of training could do anything approaching this. I was wrong.
They train it with millions (billions?) of images. It's not smart. It works because of how much processing they throw at it, and how narrow a task it's performing. Not saying these things aren't great. They could be quite useful for various tasks.
If you know how to maximize performance of AI algorithms, and you treat yourself as an AI algorithm and try to optimize yourself at a task as if you were an AI algorithm, you can have insane quick results.
I’ve done it to myself it table top puzzle solving, bowling and ice skating.
Ie by game two in bowling 80% of my rounds will be strikes or spares. My first game will look like I’m retarded as I’m establishing an “experience baseline” for the mind to try to make small deviations in the perceived correct directions.
I follow the same motions down to being ceremonial. Feet in the same exact starting positions, same exact gait through the motions etc.
That guy reminds me of how I would imagine Chester the Molester would sound.
LOL, they should be digging up his basement for his victims.
Maybe your algorithm is shitty.
(post is archived)